W3C home > Mailing lists > Public > public-ws-policy@w3.org > September 2006

NEW ISSUE: New Attribute keyword to identify 'local' policies #3721

From: William Henry <William.Henry@iona.com>
Date: Wed, 13 Sep 2006 11:42:28 -0600
To: public-ws-policy@w3.org
Message-Id: <268B62A2-BB54-4424-B23A-A04194B8DE59@iona.com>

Title: New Attribute keyword to identify 'local' policies

As WS-Policy becomes more popular in use, policies not related to
consumer/provider interaction are being defined by implementors. e.g. a
provider processes use of caching data, or a consumers private identity
management information. Though such policies should not be used in WSDL
contract it is likely that such polices could make themselves visible  
they are not to be used through attachment mechanisms (external XML  
files or
UDDI etc.)

I understand that his issue may have been raised before and the  
argument was
that it was out of scope and that domains are responsible for  
defining such
attributes. (e.g. WSDM) However this puts a burden on consumers to  
certain domain specifications or certain proprietary implementer  
policies. A
service might be deemed unusable just because consumer doesn't  
understand some
policy that is actually just a configuration policy for a local server.

A more consistent and also efficient mechanism is required. Having a  
e.g.  wsp:local (or wsp:providerOnly, wsp:consumerOnly) allows  
consumers to
ignore such policies.

What an implementor does inside that policy then is up to them and is
"invisible" to the consumer of the the policy as it will be ignored.

So though it seems like it is not clear that we should do this in the  
it does allow a mechanism to make consumer/provider policies clearer  
- i.e.
anything tagged with this attribute (e.g. wsp:local) can be ignored for
consumer/provider interaction. Pushing it out to the domain  
specifications can
leave a lot of ambiguity and therefore could effect our charter.

Proposal Description
Introduce the wsp:local attribute and explain that policies with this  
do not effect consumer/provider interaction and should be ignored from
calculating assertions for such interaction.
Received on Wednesday, 13 September 2006 17:42:40 UTC

This archive was generated by hypermail 2.4.0 : Friday, 17 January 2020 19:33:15 UTC